Merge pull request #83 from datajoint/stage

Fix regression behavior with curley bracket parsing

src/mym.cpp

@@ -399,7 +399,7 @@ void mexFunction(int nlhs, mxArray*plhs[], int nrhs, const mxArray*prhs[]) {
     /*********************************************************************/
     //  Parse the result based on the first argument
     enum querytype { OPEN, CLOSE, CLOSE_ALL, USE, STATUS, CMD, SERIALIZE, DESERIALIZE, VERSION } q;
-    char*query = NULL;
+    char* query = NULL;
     if (nrhs<=jarg)
         q = STATUS;
     else {
@@ -817,41 +817,51 @@ void mexFunction(int nlhs, mxArray*plhs[], int nrhs, const mxArray*prhs[]) {
             if (pcmp!=NULL)
                 mxFree(pcmp);
         }
-        // Check that no placeholders are present in the query except for \{ and \}
-        char* parsedQueryString = (char*)mxCalloc(strlen(query), sizeof(char)); // Query string with escape character removed
-        size_t parsedQueryStringCurrentIndex = 0;
+
+        // Remove any white spaces at the beginning (NOTE for some reason mxFree crashes if this runs before it gets to the CMD block)
+        removeWhiteSpaceAtTheBeginning(query);
+        lengthOfQuery = strlen(query);
 
-        // Loop through the query string character by character
-        for (size_t i = 0; query[i] != '\0'; i++) {
-
-            if ((query[i] == '{' || query[i] == '}') && i != 0) {
-                if (query[i - 1] != '\\') {
-                    // Curley bracket doesn't seem to be escaped thus throw and error
-                    mxFree(parsedQueryString);
-                    mexErrMsgTxt("The query contains placeholders, but no additional arguments!");
+        // Check if it is Create or Alter, if so then do curely bracket parsing
+        if (isSubstringFountAtTheBeginningCaseInsenstive(query, "CREATE") || isSubstringFountAtTheBeginningCaseInsenstive(query, "ALTER"))
+        {
+             // Check that no placeholders are present in the query except for \{ and \}
+            char* parsedQueryString = (char*)mxCalloc(strlen(query), sizeof(char)); // Query string with escape character removed
+            size_t parsedQueryStringCurrentIndex = 0;
+
+            // Loop through the query string character by character
+            for (size_t i = 0; query[i] != '\0'; i++) {
+
+                if ((query[i] == '{' || query[i] == '}') && i != 0) {
+                    if (query[i - 1] != '\\') {
+                        // Curley bracket doesn't seem to be escaped thus throw and error
+                        mxFree(parsedQueryString);
+                        mexErrMsgTxt("The query contains placeholders, but no additional arguments!");
+                    }
+                    else {
+                        // Valid curley bracket thus add it to the parsedQueryString with the \ removed
+                        parsedQueryStringCurrentIndex--;
+                        parsedQueryString[parsedQueryStringCurrentIndex] = query[i];
+                    }
                 }
                 else {
-                    // Valid curley bracket thus add it to the parsedQueryString with the \ removed
-                    parsedQueryStringCurrentIndex--;
+                    // Non curley bracket thus add it to the parsedQueryString
                     parsedQueryString[parsedQueryStringCurrentIndex] = query[i];
                 }
+
+                // Increment the currentIndex counter
+                parsedQueryStringCurrentIndex++;
             }
-            else {
-                // Non curley bracket thus add it to the parsedQueryString
-                parsedQueryString[parsedQueryStringCurrentIndex] = query[i];
-            }
+            // Add ending character
+            parsedQueryString[parsedQueryStringCurrentIndex] = '\0';
 
-            // Increment the currentIndex counter
-            parsedQueryStringCurrentIndex++;
+            // Update the query string
+            mxFree(query); // Free up the old string allocation
+            query = parsedQueryString;
+            lengthOfQuery = strlen(query); // Update the length of the query
+
         }
-        // Add ending character
-        parsedQueryString[parsedQueryStringCurrentIndex] = '\0';
-
-        // Update the query string
-        mxFree(query); // Free up the old string allocation
-        query = parsedQueryString;
-        lengthOfQuery = strlen(query); // Update the length of the query
-
+
         // Process flags
         for (int i=nrhs-nb_flags; i < nrhs; ++i) {
             if  (strcasecmp(getstring(prhs[i]), ML_FLAG_BIGINT_TO_DOUBLE)==0) {
@@ -1179,6 +1189,58 @@ void mexFunction(int nlhs, mxArray*plhs[], int nrhs, const mxArray*prhs[]) {
     }
 }
 
+void removeWhiteSpaceAtTheBeginning(char* string) 
+{
+    // Do a quick check at the start to see if any work needs to be done
+    if (!strlen(string) || !isspace(string[0])) 
+    {
+        // The first character is not an empty space thus just return the orignal string
+        return;
+    }
+
+    // There are some white spaces that needs to be removed
+    size_t currentStringIndex = 1; // Offset of 1 due to the check at the beginning
+
+    // Loop through the string until a none whitespace character is found
+    for (; string[currentStringIndex] != '\0'; currentStringIndex++) 
+    {
+        if (string[currentStringIndex] != ' ') 
+        {
+            break;
+        }
+    }
+
+    char* sanitizeString = (char*)mxCalloc(strlen(string), sizeof(char));
+    size_t sanitizeStringIndex = 0;
+    // Copy the rest of the string over
+    for (; string[currentStringIndex] != '\0'; currentStringIndex++) 
+    {
+        sanitizeString[sanitizeStringIndex] = string[currentStringIndex];
+        sanitizeStringIndex++;
+    }
+
+    // Replace the string
+    mxFree(string); // Free up the old string allocation
+    string = sanitizeString;
+}
+
+/**
+ * Helper function for checking if string subString is found at the start of string sourceString
+ * @param sourceString Source string to find subString at the start
+ * @param subString String target to look at the beginning of sourceString
+ * @returns Boolean answer if subString was found at the start of sourceString
+ */
+bool isSubstringFountAtTheBeginningCaseInsenstive(const char* sourceString, const char* subString)
+{
+    for (size_t i = 0; subString[i] != '\0'; i++) 
+    {
+        if (tolower(sourceString[i]) != tolower(subString[i]))
+        {
+            return false;
+        }
+    }
+    return true;
+}
 
 /*************************************SERIALIZE******************************************/
 // Serialize a structure array

src/mym.h

@@ -34,7 +34,7 @@
 // mym version information
 #define MYM_VERSION_MAJOR 2
 #define MYM_VERSION_MINOR 8
-#define MYM_VERSION_BUGFIX 1
+#define MYM_VERSION_BUGFIX 2
 
 
 // some local defintion
@@ -48,6 +48,7 @@
 #include <mex.h>  //  Definitions for Matlab API
 #include <zlib.h>
 #include <math.h>
+#include <ctype.h>
 #include "matrix.h"
 
 // We need a platform- and compiler-independent (rofl) fixed size 64 bit integer
@@ -105,6 +106,8 @@ mxArray* deserializeCell(const char* rpSerial, const size_t rlength);
 int file_length(FILE *f); // get the size of a file in byte
 unsigned long min_mysql_escape(char* rpout, const char* rpin, const unsigned long nin);
 void safe_read_64uint(mwSize* dst, _uint64* src, size_t n);
+bool isSubstringFountAtTheBeginningCaseInsenstive(const char* sourceString, const char* subString);
+void removeWhiteSpaceAtTheBeginning(char* string);
 //void safe_read_64uint(mwSize* dst, unsigned __int64* src, size_t n);
 
 /**********************************************************************

tests/TestDeclare.m

@@ -10,14 +10,32 @@ function TestDeclare_delcareTable(testCase)
                 'false');
 
             mym(conn1, ['create database `' testCase.PREFIX '_declare`']);
-            mym(['create table {S}(id int, data varchar(30) default null) COMMENT "\{username\}_\{subject_nickname\}"'], ...
+            mym([' create table {S}(id int, data varchar(30) default null) ' ...
+                 'COMMENT "\{username\}_\{subject_nickname\}"'], ...
                 sprintf('`%s_%s`.`%s`', testCase.PREFIX, 'declare', 'test_table'));
-            query_string = mym(sprintf('SHOW CREATE TABLE `%s_%s`.`%s`', testCase.PREFIX, 'declare', 'test_table'));
 
-            if ~strfind(query_string.('Create Table'){1}, 'COMMENT=''{username}_{subject_nickname}''')
+            % Verify that comment got inserted correctly
+            query_string = mym(sprintf( ...
+                'SHOW CREATE TABLE `%s_%s`.`%s`', ...
+                testCase.PREFIX, ...
+                'declare', ...
+                'test_table'...
+                ));
+            if ~strfind(query_string.(...
+                    'Create Table'){1}, 'COMMENT=''{username}_{subject_nickname}''' ...
+                    )
                 throw(MException('TestDeclare:invalidDeclareTable',...
                     'Table comments got inserted incorrectly'));
             end
+
+            % Test if the curley bracket parsing mess with other data that it not suppose to
+            mym(sprintf( ...
+                'INSERT INTO `%s_%s`.`%s` (`id`, `data`) VALUES (0, ''}hello{'')', ...
+                testCase.PREFIX, ...
+                'declare', ...
+                'test_table' ...
+                ));
+
             mym(conn1, ['drop database `' testCase.PREFIX '_declare`']);
             mym(conn1, 'close');
         end